arbitrage Archives - Coinscious

Coinscious Labs Arbitrage

Coinscious Lab: Arbitrage

By | Coinscious Lab | No Comments

The Challenge of Arbitrage in Crypto Markets

People get excited when they hear about arbitrage opportunities in the crypto market. That excitement is driven by frequently seeing big price differences between exchanges. If you have coding skills, you might think taking advantage of those price differences using arbitrage will be a piece of cake.

The real challenges come as you write, test, and run your code against exchange accounts. Here are just a few of the challenges that come from trying to execute an arbitrage strategy in the crypto market.

Simulation Accuracy

Most developers use simulations to verify their ideas and coding. However, depending on how you execute your simulation, your simulation results might be quite different from your real-world  results. For example, if your simulation is based on the order book (open buy and sell orders), your simulation result might be totally misleading.

Placing Orders

When you see price differences between exchanges, do you check price differences in the order book as well? Do you know if your orders will actually be fulfilled or the timing of those orders?

Capturing Future Opportunities

After you make some arbitrage transactions, you might find one exchange price is always higher than another. Opportunities to take advantage of arbitrage may decrease. How will you continue to profit with limited funds and fewer opportunities?

Technical Challenges

The crypto exchange is not like the stock exchange. Most crypto exchange APIs are not designed for high-frequency trading. Some APIs are poorly designed and unstable. The exchange may go down. The API server could stop responding. Perhaps internal logic or a limit is changed on the exchange side. There are many scenarios where API calls fail. It might be harder than you think to execute a robust arbitrage program without human interaction.

Weighing the Risks

Transaction fees are not cheap in crypto exchanges. If you have transaction success in one exchange, but failed transactions in another, you could lose money very quickly.

Profit Expectations

Of course, everyone would like to know if profit expectations through arbitrage are real and reliable. Many variables come into play when answering this question.

Choosing an Arbitrage Coding Strategy

In addition to two-point arbitrage, you also might hear about triangle or multiple-point arbitrage. Do you understand which strategy is best for you? Does fund size affect your arbitrage strategy? What is the proper fund size you should try? Answering these questions might be harder than you thought.

Bot Parameters

The parameters used by trading bots are also important. There is no “one size fits all” bot that can simply run for you all day every day. Bot parameters have to be set, tested, optimized, and then put into practice in the real world.

Experiment Goal

In this lab experiment, we conducted an arbitrage experiment to determine how to overcome the challenges mentioned above. This lab experiment will also verify what kind of data support is needed to maximize profit and minimize risk.

Profit Snapshots 

Coinscious Labs Arbitrage

This image shows the total amount of transactions completed per day.


Coinscious Labs Arbitrage

This image shows a breakdown of all 36 transactions from January 14. The highlighted price in the top right of the image shows that we bought 13.46 LSK at a price of 0.00194800 BTC, on Binance. Correspondingly, this transaction is shown in the Binance transaction history in the snapshot below.


Our results in the Binance transaction history


Experiment Results

We ended the month with a profit of +0.81616874 BTC (+55.34%), and +80.02477571 LSK (+8.07%).

We see our transactions in the real world, fully automated, with a dynamic configuration of different variables and parameters.

This lab experiment was conducted on several combinations of currencies (such as: BTC/LSK, BTC/USD, BTC/IOTA, BTC/ETH) and on different exchanges.

All experiments were successful. In this Coinscious Lab experiment, we looked at the results of our BTC/LSK arbitrage. For data and information on other experiments, please contact us.

Coinscious Labs Arbitrage Results

Here we see the accumulative results of arbitrage for each day during the month of January.


Future Development

Based on different experiments completed by Coinscious Lab, the Coinscious product team will create new product designs that cover the various needs in crypto arbitrage. These designs will not only help traders capture more opportunities, but also will optimize the process through simulation, AI optimization, profit estimation, and so on.


Data or Noise: Making Sense of the Cryptocurrency Trading Market

Data or Noise: Making Sense of the Crypto Trading Market

By | Cryptocurrency | No Comments

Too good to be true press reports.
Misleading ads.
Unscrupulous exchanges.
“Gold rush” fever.

These are just a few of the things that can overhype crypto-coin value. If you’re a serious investor, how do you resist the hype and rely only on data? Where do you find reliable, unbiased information to guide your trade decisions? If you’ve been in this market long, you know answering those questions is difficult. You’re left to your own ingenuity to figure it out for yourself.

Risky Business

As a cryptocurrency investor, you face risk. That risk is higher because you have no easily accessible tools to help you manage your investments. Unlike the traditional stock market, cryptocurrency investors can’t choose from a wide array of apps or measurement guides to drive their trade decisions. For active traders, it’s difficult to set up investment buy/sell parameters tied to return, liquidity, or market movement.

All investment carries risk, but in new and uncertain markets, that risk is heavier. The price volatility of cryptocurrency convinces some traders to stay away, or to only invest small sums. Others are suckered by the hype and gamble too much based on too little information.

Many factors affect coin prices—social media buzz, news reports, trade activity. It’s hard, if not impossible, for traders to calculate how each factor plays into cryptocurrency price fluctuations. To compound matters, traders are hit with lots of information at every step of the trading process. Investors find themselves making trade decisions based on scattered information from random online sources. Scouring that information takes time and adds uncertainty to the process.

How AI and Machine Learning Manage Risk

The cryptocurrency market is nuanced, complex, and volatile. As this market matures, it attracts more interested traders. That interest will create demand for tools that support both casual and professional investors at each step of the trading process—from building a portfolio to developing and improving trade strategies. Any solution must address the two core questions that traders ask: How can I understand the market to better build my portfolio? How can I manage my trade strategy?

These are also the questions that Coinscious has asked. Our team of blockchain engineers, data scientists, and AI researchers found the answers in AI and deep learning. Using deep neural networks, the Coinscious Collective™ platform has been trained to automatically recognize patterns and make predictions by observing data. The platform applies statistical machine learning algorithms to the cryptocurrency market to understand and recognize the patterns in available data.

With these technologies, the platform extracts the information needed to help traders make informed decisions. The platform’s AI and machine learning capabilities focus on areas that investors care about most: future value prediction, uncertainty estimation, market correlation detection, and monetary and coin movement patterns.

Separating Data from Noise

The goal of the Coinscious Collective™ platform is to separate signal from noise in price data, trade data, social media data, etc. We have three goals:

  • To be immune to noise as much as possible
  • To determine the noise-to-signal ratio at any given time
  • To improve the decision making process for traders and investors

Whether in the stock market or crypto market, information overload often drowns out the valuable signal. The ability to separate noise from data helps investors better evaluate market performance and create smarter strategies. In fact, studies show that what separates successful traders from all the others is the ability to distinguish data that matters from information that doesn’t.

With the explosive growth of data, information overload is a common problem and causes traders to get distracted. This creates a number of problems for serious traders:

  • The chance of discovering meaningful information is equivalent to finding a needle in an exponentially growing haystack.
  • There are limits to how much information any human can process and filter.
  • Even if you overcome the first two limits, human bias accumulates over time and affects the filtering process.

To remedy this, Coinscious leverages an AI system. Using deep learning, the platform uncovers hidden patterns. The platform achieves this through gathering a good representation of data (i.e. feature engineering), ensuring the model extracts noise while also generalizing future trends, and modeling uncertainty. It’s with this process that the Coinscious Collective™ platform limits non-essential data and delivers statistically credible information. If you’re a serious investor, you can spend your day trying to find the right information, or you can let Coinscious deliver the right information to you.

Want a deeper understanding of the technology behind the Coinscious Collective™ platform? Learn how Coinscious works.